@chetaslua mythos was made available for internal testing on February 24, 2026 i guess you did a typo like the Oceanus is better then mythos it is the same model or you mean this svg level was achieved by gemini 2.5
Oh you'd feel it it is already felt
your work is just done before and in the training data + you have tests it can try infinite time and stumble into the solution
uber Microsoft salesforces etc
try to replicate an artisictic slide from pdf to a powerpoint you would see it fail miserably
the premise is not that llms are useless but they won't achieve general intelligence notice the article i'm replying to
these ai use cases existed for years before llms the fact that you are mentioning them proves there are no big llm use cases YET
ad reduction is classic ML (reduction gradient boositng)
protein folding are graph neural nets
fraud detection again classical ML gradient boosted trees
the fact that you don't know what you are talking about + the nice straw man there as if i only mentioned uber
i will give you one more example salesforce has been rehiring after it reduced headcount thinking ai (llms) would replace so now cope away
@AnthropicAI bro you might wanna hop off ai coding
I'm genuinely impressed no one found a bug in your internal systems and leaked the entire ip with that shity code quality of yours
@AnthropicAI bro you might wanna hop off ai coding
I'm genuinely impressed no one found a bug in your internal systems and leaked the entire ip with that shity code quality of yours
Our internal data shows Claude is accelerating AI developmentโa possible path to recursive self-improvement, or AI autonomously building a more capable successor.
Itโs happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx
@SemiAnalysis_ A quick question if LLMS are not the way and another ai paradigm emerges and we need general computing are these chips useful and general purpose like gpus
you know @LinusTech has fallen off when they didn't grab that windows build with drivers and ran benchmarks with DGX spark (contains the same chip as rtx spark)